|
In order to realize the detection of targets in near-infrared remote sensing images, a target detection system based on traditional algorithms and random rough surface theory (Dot-curve system) is established. The traditional algorithm and Dot-curve system algorithm used in the system are studied. First, according to the target's morphological characteristics, gray-scale characteristics, feature invariance, signal scale invariance and other principle characteristics, the traditional target detection algorithm is introduced. Then, the structure diagram of the algorithm in this paper is given, the definition and principle of Dot-curve system algorithm are introduced in detail, and the nonlinear fitting algorithm used in this paper is introduced and explained. After this, based on the analysis and comparison of the target skeleton's main axis gray average, minor axis gray average, main axis gray variance, minor axis gray variance, and angle of the minor axis gray slope, we verified Dot-curve system algorithm by simulation. Finally, the output shows the Dot point of the target and all the main and minor axes, and the gray curve of the main and minor axes after fitting is compared and analyzed. The experimental results show that the simulation fitting effect is good, and the ship target and wake can be detected intuitively. Because of it completely relies on the theory of graphic feature analysis, this method has the advantages of high stability and strong anti-interference ability.
|
